Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: [OPEN SOURCE] corticod (neural multimedia cocec) (Read 2460 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

[OPEN SOURCE] corticod (neural multimedia cocec)

Publishing date: Jan 21, 2025
corticod is a biomimetic cortical coding library designed for hierarchical data processing, pattern recognition, and sequential data compression. It models a neural-inspired tree structure for adaptive learning and sparse representation of data.

  • A neural multimedia codec, currently tested for audio coding. Have comparable performance to current state-of-the-art lossy audio codecs (MP3, AAC, OPUS, OGG).

Official git: https://github.com/itu-lab/CorticalCodingCodec#readme
Hybrid Multimedia Production Suite will be a platform-indipendent open source suite for advanced audio/video contents production.
Official git: https://forart.it/HyMPS/

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #1 –
Quote from: forart.eu on
A neural multimedia codec, currently tested for audio coding. Have comparable performance to current state-of-the-art lossy audio codecs (MP3, AAC, OPUS, OGG).
nice info

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #2 –
OK. Some binaries would be nice, and info does it require GPU and which ones, if it does.
Last time I tested some AI codec, it needed nVidia GPU or abysmal encoding speed (and decoding!) on CPU.
TAPE LOADING ERROR

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #4 –
I don't know how many neural codecs we've heard of, but they are unfortunately not practical for now. Hundreds of MB model sizes, very long and expensive training processes, very slow processing times, high energy consumption....
And of course I'm not even talking about music used without any rules, permissions or rights. This is all just a tiny improvement over current audio codecs. Despite this, lossy results are always controversial. And only chip manufacturers always benefit from this situation.

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #5 –
Quote from: genuine on
And of course I'm not even talking about music used without any rules, permissions or rights.
This is getting ridiculous: what kind of permission would ever be needed to train an ENCODER?!?
Hybrid Multimedia Production Suite will be a platform-indipendent open source suite for advanced audio/video contents production.
Official git: https://forart.it/HyMPS/

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #6 –
They should provide a list of the music used in education. And they must explain how they obtained them. But they can't do that.
These systems operate by exploiting countless works of countless artists without paying anything. You may support such works, but I cannot respect it.

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #7 –
Quote from: genuine on
They should provide a list of the music used in education.
...why ?

Quote from: genuine on
These systems operate by exploiting countless works of countless artists without paying anything.
I can't exactly understand the connection between ARTISTS' rights and audio encoding...
...or (maybe) are you confusing with "AI" music generators, are you?

Anyway, I do believe that (actually) NeuralNetworks should be better used for "decoding" lossy audio to near-lossless...
Hybrid Multimedia Production Suite will be a platform-indipendent open source suite for advanced audio/video contents production.
Official git: https://forart.it/HyMPS/

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #8 –
We upset each other unnecessarily over such codecs that have no practical equivalent in real life.

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #9 –
Quote from: genuine on
We upset each other unnecessarily over such codecs that have no practical equivalent in real life.
FlowDec and AudioDec - by Meta - has been released (both under CC BY-NC License)...

Check their demo pages:
FlowDec - https://sp-uhh.github.io/FlowDec/
AudioDec - https://bigpon.github.io/AudioDec_demo/

Last but not least, I'm not upset at all.
Anyway I do apreciate CONSTRUCTIVE confrontation.
Hybrid Multimedia Production Suite will be a platform-indipendent open source suite for advanced audio/video contents production.
Official git: https://forart.it/HyMPS/

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #10 –
FlowDec 1.12 gb and AudioDec 648 mb. Unfortunately, I don't have enough time or energy to deal with them. There is no need for such brute force things for the sake of a very small gain from lossy compression.
Using executable files of several hundred kb with traditional codecs, similar quality results can be achieved in comic times. In other words, there is no longer a significant difference between 4 kbps and 40 kbps in terms of disk and data transfer today.

Re: [OPEN SOURCE] corticod (neural multimedia cocec)

Reply #11 –
Quote from: genuine on
These systems operate by exploiting countless works of countless artists without paying anything.
in all fairness, many of the legal avenues in which these works are operated under do not facilitate giving your money directly to the artist you "love", it goes to copyright owners instead and only that.
any other "revolutionary" music distribution service that claims otherwise is lying to you.

and with regards to training on unauthorized works, there will come a point where the ones who led the way to AI-based tech will have to provide training data themselves, but this time it'll be done from their own talents only instead of talented people who had no such input in such tech existing but have made clear that they'll reject them with extreme prejudice if they even so much as attempt to replicate their talents.

alas, last time anyone even remotely entertained the thought of bringing such issues to this very community, john33 just went and told them off entirely.